MCPcopy Create free account
hub / github.com/dotchen/LearningByCheating / get_control

Function get_control

training/train_image_phase2.py:45–58  ·  view source on GitHub ↗

A learning schedule to choose between agent control and teacher control lim_{episode->inf} P(control=agent_control|episode) = 1 to make sure it converges

(agent_control, teacher_control, episode, beta=0.95)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 1

rolloutFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected